| Message ID | 20220215230737.1870630-2-djrscally@gmail.com |
|---|---|
| State | New |
| Headers | show |
| Series | [01/10] media: uapi: Add IPU3 packed Y10 format | expand |
Le mardi 15 février 2022 à 23:07 +0000, Daniel Scally a écrit : > Some platforms with an Intel IPU3 have an IR sensor producing 10 bit > greyscale format data that is transmitted over a CSI-2 bus to a CIO2 > device - this packs the data into 32 bytes per 25 pixels. Detail that > format. > > Signed-off-by: Daniel Scally <djrscally@gmail.com> > --- > .../userspace-api/media/v4l/pixfmt-yuv-luma.rst | 14 +++++++++++++- > drivers/media/v4l2-core/v4l2-ioctl.c | 1 + > include/uapi/linux/videodev2.h | 1 + > 3 files changed, 15 insertions(+), 1 deletion(-) > > diff --git a/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst b/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst > index 8ebd58c3588f..5465ce3bb533 100644 > --- a/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst > +++ b/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst > @@ -48,6 +48,17 @@ are often referred to as greyscale formats. > - ... > - ... > > + * .. _V4L2-PIX-FMT-IPU3-Y10: > + > + - ``V4L2_PIX_FMT_IPU3_Y10`` > + - 'ip3y' > + > + - Y'\ :sub:`0`\ [7:0] > + - Y'\ :sub:`1`\ [5:0] Y'\ :sub:`0`\ [9:8] > + - Y'\ :sub:`2`\ [3:0] Y'\ :sub:`1`\ [9:6] > + - Y'\ :sub:`3`\ [1:0] Y'\ :sub:`2`\ [9:4] > + - Y'\ :sub:`3`\ [9:2] > + > * .. _V4L2-PIX-FMT-Y10: > > - ``V4L2_PIX_FMT_Y10`` > @@ -133,4 +144,5 @@ are often referred to as greyscale formats. > > For the Y16 and Y16_BE formats, the actual sampling precision may be lower > than 16 bits. For example, 10 bits per pixel uses values in the range 0 to > - 1023. > + 1023. For the ip3y format 25 pixels are packed into 32 bytes, which leaves nit: ip3y-> IPU3_Y10, to be consistent with previous paragraph not using fourcc. I don't have very strong preference, but this could have been sorted into vendor formats, as its specific to a HW design, unlike MIPI which is a HW standard. In any case, with the nit fixed, you can add my: Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com> cheers, Nicolas > + the 6 most significant bits of the last byte padded with 0. > diff --git a/drivers/media/v4l2-core/v4l2-ioctl.c b/drivers/media/v4l2-core/v4l2-ioctl.c > index 642cb90f457c..89691bbb372d 100644 > --- a/drivers/media/v4l2-core/v4l2-ioctl.c > +++ b/drivers/media/v4l2-core/v4l2-ioctl.c > @@ -1265,6 +1265,7 @@ static void v4l_fill_fmtdesc(struct v4l2_fmtdesc *fmt) > case V4L2_PIX_FMT_Y16_BE: descr = "16-bit Greyscale BE"; break; > case V4L2_PIX_FMT_Y10BPACK: descr = "10-bit Greyscale (Packed)"; break; > case V4L2_PIX_FMT_Y10P: descr = "10-bit Greyscale (MIPI Packed)"; break; > + case V4L2_PIX_FMT_IPU3_Y10: descr = "10-bit greyscale (IPU3 Packed)"; break; > case V4L2_PIX_FMT_Y8I: descr = "Interleaved 8-bit Greyscale"; break; > case V4L2_PIX_FMT_Y12I: descr = "Interleaved 12-bit Greyscale"; break; > case V4L2_PIX_FMT_Z16: descr = "16-bit Depth"; break; > diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h > index df8b9c486ba1..b378c7e37eac 100644 > --- a/include/uapi/linux/videodev2.h > +++ b/include/uapi/linux/videodev2.h > @@ -569,6 +569,7 @@ struct v4l2_pix_format { > /* Grey bit-packed formats */ > #define V4L2_PIX_FMT_Y10BPACK v4l2_fourcc('Y', '1', '0', 'B') /* 10 Greyscale bit-packed */ > #define V4L2_PIX_FMT_Y10P v4l2_fourcc('Y', '1', '0', 'P') /* 10 Greyscale, MIPI RAW10 packed */ > +#define V4L2_PIX_FMT_IPU3_Y10 v4l2_fourcc('i', 'p', '3', 'y') /* IPU3 packed 10-bit greyscale */ > > /* Palette formats */ > #define V4L2_PIX_FMT_PAL8 v4l2_fourcc('P', 'A', 'L', '8') /* 8 8-bit palette */
Hi Nicolas On 16/02/2022 13:28, Nicolas Dufresne wrote: > Le mardi 15 février 2022 à 23:07 +0000, Daniel Scally a écrit : >> Some platforms with an Intel IPU3 have an IR sensor producing 10 bit >> greyscale format data that is transmitted over a CSI-2 bus to a CIO2 >> device - this packs the data into 32 bytes per 25 pixels. Detail that >> format. >> >> Signed-off-by: Daniel Scally <djrscally@gmail.com> >> --- >> .../userspace-api/media/v4l/pixfmt-yuv-luma.rst | 14 +++++++++++++- >> drivers/media/v4l2-core/v4l2-ioctl.c | 1 + >> include/uapi/linux/videodev2.h | 1 + >> 3 files changed, 15 insertions(+), 1 deletion(-) >> >> diff --git a/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst b/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst >> index 8ebd58c3588f..5465ce3bb533 100644 >> --- a/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst >> +++ b/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst >> @@ -48,6 +48,17 @@ are often referred to as greyscale formats. >> - ... >> - ... >> >> + * .. _V4L2-PIX-FMT-IPU3-Y10: >> + >> + - ``V4L2_PIX_FMT_IPU3_Y10`` >> + - 'ip3y' >> + >> + - Y'\ :sub:`0`\ [7:0] >> + - Y'\ :sub:`1`\ [5:0] Y'\ :sub:`0`\ [9:8] >> + - Y'\ :sub:`2`\ [3:0] Y'\ :sub:`1`\ [9:6] >> + - Y'\ :sub:`3`\ [1:0] Y'\ :sub:`2`\ [9:4] >> + - Y'\ :sub:`3`\ [9:2] >> + >> * .. _V4L2-PIX-FMT-Y10: >> >> - ``V4L2_PIX_FMT_Y10`` >> @@ -133,4 +144,5 @@ are often referred to as greyscale formats. >> >> For the Y16 and Y16_BE formats, the actual sampling precision may be lower >> than 16 bits. For example, 10 bits per pixel uses values in the range 0 to >> - 1023. >> + 1023. For the ip3y format 25 pixels are packed into 32 bytes, which leaves > nit: ip3y-> IPU3_Y10, to be consistent with previous paragraph not using fourcc. Ack > > I don't have very strong preference, but this could have been sorted into vendor > formats, as its specific to a HW design, unlike MIPI which is a HW standard. Yeah I vacillated between putting it in here and with the other IPU3 formats [1], but because those are in the "bayer formats" section it sorta stuck out so I eventually settled on sticking it with greyscale...but maybe there needs to be another section for "Hardware Specific Formats" or something instead of either option. [1] https://linuxtv.org/downloads/v4l-dvb-apis/userspace-api/v4l/pixfmt-srggb10-ipu3.html > > In any case, with the nit fixed, you can add my: > > Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com> Thank you Dan > > cheers, > Nicolas > >> + the 6 most significant bits of the last byte padded with 0. >> diff --git a/drivers/media/v4l2-core/v4l2-ioctl.c b/drivers/media/v4l2-core/v4l2-ioctl.c >> index 642cb90f457c..89691bbb372d 100644 >> --- a/drivers/media/v4l2-core/v4l2-ioctl.c >> +++ b/drivers/media/v4l2-core/v4l2-ioctl.c >> @@ -1265,6 +1265,7 @@ static void v4l_fill_fmtdesc(struct v4l2_fmtdesc *fmt) >> case V4L2_PIX_FMT_Y16_BE: descr = "16-bit Greyscale BE"; break; >> case V4L2_PIX_FMT_Y10BPACK: descr = "10-bit Greyscale (Packed)"; break; >> case V4L2_PIX_FMT_Y10P: descr = "10-bit Greyscale (MIPI Packed)"; break; >> + case V4L2_PIX_FMT_IPU3_Y10: descr = "10-bit greyscale (IPU3 Packed)"; break; >> case V4L2_PIX_FMT_Y8I: descr = "Interleaved 8-bit Greyscale"; break; >> case V4L2_PIX_FMT_Y12I: descr = "Interleaved 12-bit Greyscale"; break; >> case V4L2_PIX_FMT_Z16: descr = "16-bit Depth"; break; >> diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h >> index df8b9c486ba1..b378c7e37eac 100644 >> --- a/include/uapi/linux/videodev2.h >> +++ b/include/uapi/linux/videodev2.h >> @@ -569,6 +569,7 @@ struct v4l2_pix_format { >> /* Grey bit-packed formats */ >> #define V4L2_PIX_FMT_Y10BPACK v4l2_fourcc('Y', '1', '0', 'B') /* 10 Greyscale bit-packed */ >> #define V4L2_PIX_FMT_Y10P v4l2_fourcc('Y', '1', '0', 'P') /* 10 Greyscale, MIPI RAW10 packed */ >> +#define V4L2_PIX_FMT_IPU3_Y10 v4l2_fourcc('i', 'p', '3', 'y') /* IPU3 packed 10-bit greyscale */ >> >> /* Palette formats */ >> #define V4L2_PIX_FMT_PAL8 v4l2_fourcc('P', 'A', 'L', '8') /* 8 8-bit palette */
diff --git a/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst b/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst index 8ebd58c3588f..5465ce3bb533 100644 --- a/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst +++ b/Documentation/userspace-api/media/v4l/pixfmt-yuv-luma.rst @@ -48,6 +48,17 @@ are often referred to as greyscale formats. - ... - ... + * .. _V4L2-PIX-FMT-IPU3-Y10: + + - ``V4L2_PIX_FMT_IPU3_Y10`` + - 'ip3y' + + - Y'\ :sub:`0`\ [7:0] + - Y'\ :sub:`1`\ [5:0] Y'\ :sub:`0`\ [9:8] + - Y'\ :sub:`2`\ [3:0] Y'\ :sub:`1`\ [9:6] + - Y'\ :sub:`3`\ [1:0] Y'\ :sub:`2`\ [9:4] + - Y'\ :sub:`3`\ [9:2] + * .. _V4L2-PIX-FMT-Y10: - ``V4L2_PIX_FMT_Y10`` @@ -133,4 +144,5 @@ are often referred to as greyscale formats. For the Y16 and Y16_BE formats, the actual sampling precision may be lower than 16 bits. For example, 10 bits per pixel uses values in the range 0 to - 1023. + 1023. For the ip3y format 25 pixels are packed into 32 bytes, which leaves + the 6 most significant bits of the last byte padded with 0. diff --git a/drivers/media/v4l2-core/v4l2-ioctl.c b/drivers/media/v4l2-core/v4l2-ioctl.c index 642cb90f457c..89691bbb372d 100644 --- a/drivers/media/v4l2-core/v4l2-ioctl.c +++ b/drivers/media/v4l2-core/v4l2-ioctl.c @@ -1265,6 +1265,7 @@ static void v4l_fill_fmtdesc(struct v4l2_fmtdesc *fmt) case V4L2_PIX_FMT_Y16_BE: descr = "16-bit Greyscale BE"; break; case V4L2_PIX_FMT_Y10BPACK: descr = "10-bit Greyscale (Packed)"; break; case V4L2_PIX_FMT_Y10P: descr = "10-bit Greyscale (MIPI Packed)"; break; + case V4L2_PIX_FMT_IPU3_Y10: descr = "10-bit greyscale (IPU3 Packed)"; break; case V4L2_PIX_FMT_Y8I: descr = "Interleaved 8-bit Greyscale"; break; case V4L2_PIX_FMT_Y12I: descr = "Interleaved 12-bit Greyscale"; break; case V4L2_PIX_FMT_Z16: descr = "16-bit Depth"; break; diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h index df8b9c486ba1..b378c7e37eac 100644 --- a/include/uapi/linux/videodev2.h +++ b/include/uapi/linux/videodev2.h @@ -569,6 +569,7 @@ struct v4l2_pix_format { /* Grey bit-packed formats */ #define V4L2_PIX_FMT_Y10BPACK v4l2_fourcc('Y', '1', '0', 'B') /* 10 Greyscale bit-packed */ #define V4L2_PIX_FMT_Y10P v4l2_fourcc('Y', '1', '0', 'P') /* 10 Greyscale, MIPI RAW10 packed */ +#define V4L2_PIX_FMT_IPU3_Y10 v4l2_fourcc('i', 'p', '3', 'y') /* IPU3 packed 10-bit greyscale */ /* Palette formats */ #define V4L2_PIX_FMT_PAL8 v4l2_fourcc('P', 'A', 'L', '8') /* 8 8-bit palette */
Some platforms with an Intel IPU3 have an IR sensor producing 10 bit greyscale format data that is transmitted over a CSI-2 bus to a CIO2 device - this packs the data into 32 bytes per 25 pixels. Detail that format. Signed-off-by: Daniel Scally <djrscally@gmail.com> --- .../userspace-api/media/v4l/pixfmt-yuv-luma.rst | 14 +++++++++++++- drivers/media/v4l2-core/v4l2-ioctl.c | 1 + include/uapi/linux/videodev2.h | 1 + 3 files changed, 15 insertions(+), 1 deletion(-)